The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from .docx files, creating a potential surface for indirect instructions.
Ingestion points: scripts/read_docx.py and scripts/convert_docx.py read content from user-specified file paths.
Boundary markers: The skill instructions lack explicit delimiters or warnings to ignore commands within the document text.
Capability inventory: The skill is capable of reading local files and writing converted text files to the /tmp directory.
Sanitization: No content filtering is applied to the extracted text before it is returned to the agent.

docx-processor